• Definitions, motives, history, spectrum
• Current practices
• A selection of tools to improve reproducibility
• Challenges, standards & our role in the future of reproducible research

Slides based on Ben Marwick's presentation to the UW Center for Statistics and Social Sciences (12 March 2014) (OrcID)

"The goal of reproducible research is to tie specific instructions to data analysis and experimental data so that scholarship can be recreated, better understood and verified." Max Kuhn, CRAN Task View: Reproducible Research

• Mathematics (400 BC?)
• Write scientific paper, Galileo, Pasteur, etc. (1660s?)
• Publish a pidgin algorithm and describe simulation datasets (1950s?)
• Sell magtape of code and data (1970s?)
• Place idiosyncratic dataset & software at website (1990s?)
• Publish datasets and scripts at website, eg. biology, political science, genetics, statistics (2000s?)
• Hosted integrated code and data (2020s?)

Gavish & Gonoho AAAS 2011, Oxberry 2013

"An article about computational result is advertising, not scholarship. The actual scholarship is the full software environment, code and data, that produced the result." Claerbout and Karrenbach, Proceedings of the 62nd Annual International Meeting of the Society of Exploration Geophysics. 1992

• Verification & Reliability: Find and fix bugs. Today's results == tomorrow's.
• Transparency: increased citation count, broader impact, improved institutional memory
• Efficiency: Reduces duplication of effort. Payoff in the (not so) long run
• Flexibility: When you don’t 'point-and-click' you gain many new analytic options.

Technical

• Classified/sensitive/big data
• Nondisclosure agreements & intellectual property
• Software licensing issues
• Competition
• Neither necessary nor sufficient for correctness (but essential for dispute resolution)

Cultural & personal

• Few researchers follow even minimal reproducibility standards.
• Few expect reproducibility
• No uniform standards of reproducibility, so no established user base
• Inertia & embarassment

Peng 2011, Science 334(6060) pp. 1226-1227

http://www.stodden.net/AMP2011/slides/pengslides.pdf

• Lots of human effort for tedious & time-wasting tasks
• Error-prone due to manual & ad hoc data handling (column and row offsets are common)
• Difficult to record - hard to reconstruct a 'click history'
• Tiny changes in data or method require extensive reworking efforts

• Claimed high debt-to-GDP ratios led to low GDP growth
• Threshold to growth at a debt-to-GDP ratio of >90%
• Substantial popular impact on autsterity politics

A coding error in their Excel spreadsheet sliced several countries out of the data set…. The Economist

In May 2015 Science retracted a study of how canvassers can sway people's opinions about gay marriage published just 5 months ago.

Original survey data not made available for independent reproduction of results. + Survey incentives misrepresented. + Sponsorship statement false. - Science Editor-in-Chief Marcia McNutt

Two Berkeley grad students who attempted to replicate the study quickly discovered that the data must have been faked.

Source

2013 Seizure study retracted after authors realize data got "terribly mixed"

The article has been retracted at the request of the authors. After carefully re-examining the data presented in the article, they identified that data of two different hospitals got terribly mixed. The published results cannot be reproduced in accordance with scientific and clinical correctness. Authors of Low Dose Lidocaine for Refractory Seizures in Preterm Neonates

Source

Authors informed the journal that the merge of lab results and other survey data used in the paper resulted in an error regarding the identification codes. Results of the analyses were based on the data set in which this error occurred. Further analyses established the results reported in this manuscript and interpretation of the data are not correct.

"Lower levels of CSF IL-6 were associated with current depression and with future depression […]" Original conclusion

"Higher levels of CSF IL-6 and IL-8 were associated with current depression […]" Revised conclusion

Source

"Instead of imagining that our main task is to instruct a computer what to do, let us concentrate rather on explaining to humans what we want the computer to do." Donald E. Knuth, Literate Programming, 1984

For example… Let's calculate the current time in R.

time <- format(Sys.time(), "%a %d %b %X %Y")

The text and R code are interwoven in the output:

The time is `r time`

The time is Mon 21 Sep 15:05:21 2015

Payoffs

• Eases collaboration
• Can track changes in any file type (ideally plain text)
• Can revert file to any point in its tracked history

Costs - Learning curve

Payoffs - Free space for hosting (and paid options) - Assignment of persistent DOIs - Tracking citation metrics

Costs - Sometimes license restrictions (CC-BY & CC0) - Limited or no private storage space

• Good: Use code with an integrated development environment (IDE). Minimize pointing and clicking (RStudio)
• Better: Use version control. Help yourself keep track of changes, fix bugs and improve project management (RStudio & Git & GitHub or BitBucket)
• Best: Use embedded narrative and code to explicitly link code, text and data, save yourself time, save reviewers time, improve your code. (RStudio & Git & GitHub or BitBucket & rmarkdown & knitr & data repository)

Stodden (IASSIST 2010) sampled American academics registered at the Machine Learning conference NIPS (134 responses from 593 requests (23%). Red = communitarian norms, Blue = private incentives

Stodden (IASSIST 2010) sampled American academics registered at the Machine Learning conference NIPS (134 responses from 593 requests (23%). Red = communitarian norms, Blue = private incentives

Reproducible Research Standard (Stodden 2009)

• The full compendium on the internet
• Media such as text, figures, tables with Creative Commons Attribution license (CC-BY)
• Code with one of Apache 2.0, MIT, LGPL, BSD, etc.
• Original "selection and arrangement" of data with CC0 or CC-BY

Promote culture change through positive attribution

Implement mechanisms to indicate & encourage degrees of compliance (ie. clear definitions for different levels of reproducibility), cf. Stodden's:

• 'Reproducible': compendium of text-code-data online
• 'Reproduced': compendium available and independently reproduced
• 'Semi-Reproducible': when the full compendium is not released
• 'Semi-Reproduced': independent reproduction with other data
• 'Perpetually Reproducible': streaming data

• Train students by putting homework, assignments & dissertations on the reproducible research spectrum
• Publish examples of reproducible research in our field
• Request code & data when reviewing
• Submit to & review for journals that support reproducible research
• Critically review & audit data management plans in grant proposals
• Consider reproducibility wherever possible in hiring, promotion & reference letters.

1. Create a GitHub account at https://github.com/
   • This will be a public account associated with your name
   • Choose a username wisely for future use
   • Don't worry about details, you can fill them in later
2. Create a repository called demo
   • Add a brief and informative description
   • Choose "Public"
   • Check the box for "Initialize this repository with a README"
3. Click "Create Repository"

Demo drawn using materials from Dr. Çetinkaya-Rundel

1. Go to RStudio
   • File -> New Project
   • Version Control: Checkout a project from a version control repository
   • Git: Clone a project from a repository
2. Fill in the info:
   • URL: use HTTPS address
   • Create as a subdirectory of: Browse and create a new folder called GEO503 (or similar)

Live demo

Cheatsheet:

https://www.rstudio.com/wp-content/uploads/2015/02/rmarkdown-cheatsheet.pdf

File -> New File -> RMarkdown -> Document -> HTML

All R code to be run must be in a code chunk like this:

#```{r,eval=F}
CODE HERE
#```

Load these packages in a code chunk:

library(dplyr)
library(ggplot2)
library(spocc)

Do you think you should put install.packages() calls in your script?

Now use the occ() function to download all the occurrence records for the American robin (Turdus migratorius) from the Global Biodiversity Information Facility.

Licensed under CC BY-SA 3.0 via Wikimedia Commons

This can take a few seconds.

## define which species to query
sp='Turdus migratorius'

## run the query and convert to data.frame()
d = occ(query=sp, from='ebird',limit = 10000) %>% occ2df()

ggplot(d,aes(x=format(date,"%m"),y=latitude,group=1))+ geom_point()+ geom_smooth()

Update the YAML header to keep the markdown file

From this:

title: "Untitled"
author: "Adam M. Wilson"
date: "September 21, 2015"
output: html_document

To this:

title: "Demo"
author: "Adam M. Wilson"
date: "September 21, 2015"
output: 
  html_document:
      keep_md: true

1. Stage
2. Commit (with a message)
3. Push

Abandoning the habit of secrecy in favor of process transparency and peer review was the crucial step by which alchemy became chemistry.Raymond, E. S., 2004, The art of UNIX programming: Addison-Wesley.

Slides based on Ben Marwick's presentation to the UW Center for Statistics and Social Sciences (12 March 2014) (OrcID)

Licensing: * Presentation: CC-BY-3.0 * Source code: MIT

See Rpres file on github for full references and sources